The effect of asynchronous execution and message latency on Max-Sum

Roie Zivan, Omer Perry, Ben Rachmut, William Yeoh

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Max-sum is a version of belief propagation that was adapted for solving distributed constraint optimization problems (DCOPs). It has been studied theoretically and empirically, extended to versions that improve solution quality and converge rapidly, and is applicable to multiple distributed applications. The algorithm was presented both as a synchronous and an asynchronous algorithm, however, neither the differences in the performance of these two execution versions nor the implications of message latency on the two versions have been investigated to the best of our knowledge. We contribute to the body of knowledge on Max-sum by: (1) Establishing the theoretical differences between the two execution versions of the algorithm, focusing on the construction of beliefs; (2) Empirically evaluating the differences between the solutions generated by the two versions of the algorithm, with and without message latency; and (3) Establishing both theoretically and empirically the positive effect of damping on reducing the differences between the two versions. Our results indicate that in contrast to recent published results indicating the drastic effect that message latency has on distributed local search, damped Max-sum is robust to message latency.

Original languageEnglish
Title of host publication27th International Conference on Principles and Practice of Constraint Programming, CP 2021
EditorsLaurent D. Michel
PublisherSchloss Dagstuhl- Leibniz-Zentrum fur Informatik GmbH, Dagstuhl Publishing
ISBN (Electronic)9783959772112
DOIs
StatePublished - 1 Oct 2021
Event27th International Conference on Principles and Practice of Constraint Programming, CP 2021 - Virtual, Montpellier, France
Duration: 25 Oct 202129 Oct 2021

Publication series

NameLeibniz International Proceedings in Informatics, LIPIcs
Volume210
ISSN (Print)1868-8969

Conference

Conference27th International Conference on Principles and Practice of Constraint Programming, CP 2021
Country/TerritoryFrance
CityVirtual, Montpellier
Period25/10/2129/10/21

Keywords

  • Distributed constraints
  • Distributed problem solving

Fingerprint

Dive into the research topics of 'The effect of asynchronous execution and message latency on Max-Sum'. Together they form a unique fingerprint.

Cite this